GPU cloud

Cudo Compute gives you access to on-demand GPU instances with per-second billing. Cudo is dedicated to providing the best user experience, whether it is via the web console, our command line tool or our terraform provider. All this with servers hosted in data centers with ISO 27001 security and ISO 9001 quality certifications.

When you launch an instance, the GPUs are connected directly via PCI passthrough so you can get the full performance. Cudo offers NVIDIA A6000 (48GB), A5000 (24GB), and A4000 (16GB) GPUs providing a range of options to cover your workload at the lowest possible cost to you. Additionally, the NVIDIA GPUs are paired with AMD EPYC CPUs for blazing performance.

AMD

Strategic Partner

Why Cudo Compute for GPUs?

Whether your field is data science, deep learning, rendering or high-performance computing on GPU, getting started is simple. Launch your instance and pull a docker container with an NVIDIA optimised distribution of your application. See our guides for more information.

Instances are powered by KVM virtualization and offer NVIDIA’s latest generation Ampere architecture with up to 48 GB of VRAM and tensor cores for up to an order of magnitude improvement in performance with machine learning workloads. Visit our marketplace to see prices and deploy an instance.

Learn more about GPU cloud
🍪 We use cookies to make your experience better.